[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[hjrrockies]

Has anyone else encountered a displaced center of lift issue? I sometimes have aircraft that, despite being symmetrical, have a center of lift towards one side or the other. Sometimes, by only detaching and reattaching the wings, a previously stable craft gets this problem. It isn't just the marker in the SPH - it manifests itself in flight as well.

I'm not on my main machine, so I will post screenshots later if necessary.

[jrandom]

Has anyone else encountered a displaced center of lift issue? I sometimes have aircraft that, despite being symmetrical, have a center of lift towards one side or the other. Sometimes, by only detaching and reattaching the wings, a previously stable craft gets this problem. It isn't just the marker in the SPH - it manifests itself in flight as well.

I'm not on my main machine, so I will post screenshots later if necessary.

Save and re-load your craft and see if the CoL is still misplaced. I'm often hitting a weird bug where moving wings/surfaces around makes the CoL display incorrectly until I save/load it.

[hjrrockies]

Save and re-load your craft and see if the CoL is still misplaced. I'm often hitting a weird bug where moving wings/surfaces around makes the CoL display incorrectly until I save/load it.

I'll give it a shot. Thanks

[Van Disaster]

I've been getting an odd displaced CoL marker with p-wings: if I load a craft & mouseover a p-wing and then hit one of the p-wing adjust keys, even if I don't actually move anything the marker pops into a new position. Not sure if that's FAR or p-wings though.

[Zidane]

Has anyone else encountered a displaced center of lift issue? I sometimes have aircraft that, despite being symmetrical, have a center of lift towards one side or the other. Sometimes, by only detaching and reattaching the wings, a previously stable craft gets this problem. It isn't just the marker in the SPH - it manifests itself in flight as well.

I'm not on my main machine, so I will post screenshots later if necessary.

I've been getting an odd displaced CoL marker with p-wings: if I load a craft & mouseover a p-wing and then hit one of the p-wing adjust keys, even if I don't actually move anything the marker pops into a new position. Not sure if that's FAR or p-wings though.

I had that problem too, it got fixed when I updated both pwings and FAR. I think pwing had an incompatibility with the newer version of far.

From the pwings changelog:

0.9.1:

-NathanKell update! Added compatibility with latest FAR

Since Pwings doesn't have support for AVC, you don't get a warning about it.

Sorry about reposting my question but it kinda got buried. The air intakes aren't showing any drag on the right click in-flight menu or on the SPH, either open or closed. Is this working as intended?

EDIT: Thanks for the answer tetryds. Wrote it here because I didn't think it didn't warrant enough purpose for a new thread.

Edited November 8, 2014 by Zidane

[tetryds]

I had that problem too, it got fixed when I updated both pwings and FAR. I think pwing had an incompatibility with the newer version of far.

From the pwings changelog:

Since Pwings doesn't have support for AVC, you don't get a warning about it.

Sorry about reposting my question but it kinda got buried. The air intakes aren't showing any drag on the right click in-flight menu or on the SPH, either open or closed. Is this working as intended?

Yes, intakes use stock drag, IIRC.

If its not stock its at least a different way.

[ferram4]

So, just an update for people since I made a serious breakthrough on something I've wanted FAR to have for a long time. As you probably know, FAR requires wing geometry to be manually defined in a config file in order for it to handle everything properly. Further, it requires that every wing be able to be defined as a trapezoid, with the root and tip of the part being parallel. This, of course, is very limiting and frustrating, as well as a possible source of errors if someone implements it wrong.

Well, I just managed to implement an algorithm that allows FAR to determine the geometry of a wing from the mesh alone, assuming that it is completely convex in shape:

Now, this means that wings with concave portions (such as the stock old unmoving winglet and the gigantic B9 swept wings) won't be perfectly resolved with this. However, this system also allows manual definition of the shape for those wings, so stranger shapes can be accounted for.

Hopefully, more good news in the coming weeks on other stuff. FAR is probably going to get a lot more realistic, and at the same time, probably perform better.

[HoneyFox]

So, just an update for people since I made a serious breakthrough on something I've wanted FAR to have for a long time. As you probably know, FAR requires wing geometry to be manually defined in a config file in order for it to handle everything properly. Further, it requires that every wing be able to be defined as a trapezoid, with the root and tip of the part being parallel. This, of course, is very limiting and frustrating, as well as a possible source of errors if someone implements it wrong.

Well, I just managed to implement an algorithm that allows FAR to determine the geometry of a wing from the mesh alone, assuming that it is completely convex in shape:

http://i.imgur.com/WZiVG6X.png

Now, this means that wings with concave portions (such as the stock old unmoving winglet and the gigantic B9 swept wings) won't be perfectly resolved with this. However, this system also allows manual definition of the shape for those wings, so stranger shapes can be accounted for.

Hopefully, more good news in the coming weeks on other stuff. FAR is probably going to get a lot more realistic, and at the same time, probably perform better.

Oh my... finally we can have more accurate wing parameters based on mesh's vertices data directly.

One thing BTW, I hope that those flight helper can be improved a bit and I might do that on my side and send you some pull request later. Perhaps I will need some help on how to get some certain values inside FAR, hope that you can offer me some help in case.

[ThorBeorn]

Fantastic news ferram4. It'll be interesting to see how epilliptical wings behave with this. Now even more than before, there probably also be more stuff to look up/study for us users when it comes to different shapes which is very inspirational to me.

I have a feeling getting that pilot license one day will be a lot easier thanks to this

[Climberfx]

Ferram, look at this video.

I find a good showcase about the first flight simulator (Flight Unlimited, but only the original first one, in DOS) that impress me a lot.

And if you have the patience, it shows a good way to implement hot air flows on KSP to fly a glider or sort of.

[Van Disaster]

I had that problem too, it got fixed when I updated both pwings and FAR. I think pwing had an incompatibility with the newer version of far.

My install is three days old, including fresh mod downloads - both mods are definitely the latest.

Here's another odd one - the control surface on the lower wing is a spoiler only.

Spoiler removed - CoL ball is hidden in the CoM ball.

Spoiler added again.

Somewhat extreme CoL movement... ( might be biplane related, I don't know - this thing is fine in the air biplane or not, but really doesn't like getting off the runway ).

That's great news, Ferram.

Edited November 7, 2014 by Van Disaster

[Gfurst]

@ferram

Pretty cool, I've always wondered how does it handle multiple wing parts together, yet it simulates pretty well.

One of the major thing about plane design is the curvature of wing, chord incidence angle and such. What pretty much creates lift.

Normally you have stabilizers that don't create lift, but instead have its aerodynamic force to help stabilize the aircraft.

However KSP always considers wings placed on the side always as lift generators, or at least was that way.

Of course I'm sure you already know all about this stuff. I'm loosing focus here, just wondered how does FAR handles this.

@van disaster

Hum, I'm pretty sure your design is quite heavy needing quite a lot of speed to generate lift on those wings, your main wheels are pretty far back than the CoM.

This could be the reason you're not lifting off, too hard to pivot around the back wheels and not much speed to get off the ground on lift alone.

[Van Disaster]

The CoL ball being odd was biplane related, I've abandoned biplanes again, especially given I don't have a clue how the wings are interacting. I might try stepped biplane wings where they don't overlap if you look from the top, but not far enough apart that they're fully tandem ( ie both wings are on the same side of CoL ).

That plane is very heavy, the problem is acceleration, it just doesn't get up to speed by the runway end ( hence the SRBs! pointless adding all that fuel capacity if you burn it in rocket mode, although I guess I could use droptanks instead ) - a longer runway was something I was asking for ages ago . It'll pull the nose up itself at 160ish, there's a fair bit of pitch authority given the entire canard pivots. If I put the wheels near CoM then I'll get a tailstrike - I guess they could be a little further forward, although it is pitched up a couple of degrees on the tarmac anyway.

That plane also seems to have found a memory leak in something, memory use grows alarmingly :S

Edited November 7, 2014 by Van Disaster

[ThorBeorn]

ferram4: I had some massive lag/bugs where FAR threw lots of null reference exceptions and other errors. Could not decouple launch stabilisers, the FAR GUI showed zeroes on all derivatives while giving null reference exeptions etc. This was with the latest dev build from GitHub, downloaded yesterday. I would be grateful if you or someone knowledgeble could take a look at my logs.

KSP.log: https://www.dropbox.com/s/5qtyxq48z66xs54/KSP.log?dl=0

Output.log: https://www.dropbox.com/s/eogir10q15gz8ly/output_log.txt?dl=0

[legolegs]

Hello. I have an issue with FAR. Two of them actually. First issue:

Center of lift indicator moves off center if I move wing. Craft file (Stock parts+FAR): http://rghost.ru/58944964

How to reproduce:

1) Load craft

2) switch to 2x symmetry mode with angle snap off

3) pick thw wing and put it back where it was

CoL after wing moved.

Part of log after loading SPH with said craft: http://pastebin.com/JD9Fe6WJ

Full log (very long one, also I loaded old save with fragments of modded data while mods being temporarily moved away, and after that I created new clean save and made tests within it): http://rghost.ru/58945141

Second issue:

That craft I linked above rolls counterclockwise for no reason.

How to reproduce:

1) Lift up.

2) Turn off SAS, FAR utils, get rid out of trim

3) roll to any sensible position (-45...+45 degree) and leave controls (maybe you have to trim pitch)

4) watch plane slowly rolling counterclockwise.

[ThorBeorn]

legolegs:

1. That happens. Just save craft and load it and the CoP will be back were it should be.

2. This is a stock bug causing part connection on one side being slightly weaker, hence the wing flexes a tiny bit causing a little less lift. This can be fixed by attaching struts between the wing and the fuselage.

[blowfish]

So, just an update for people since I made a serious breakthrough on something I've wanted FAR to have for a long time. As you probably know, FAR requires wing geometry to be manually defined in a config file in order for it to handle everything properly. Further, it requires that every wing be able to be defined as a trapezoid, with the root and tip of the part being parallel. This, of course, is very limiting and frustrating, as well as a possible source of errors if someone implements it wrong.

Well, I just managed to implement an algorithm that allows FAR to determine the geometry of a wing from the mesh alone, assuming that it is completely convex in shape:

http://i.imgur.com/WZiVG6X.png

Now, this means that wings with concave portions (such as the stock old unmoving winglet and the gigantic B9 swept wings) won't be perfectly resolved with this. However, this system also allows manual definition of the shape for those wings, so stranger shapes can be accounted for.

Hopefully, more good news in the coming weeks on other stuff. FAR is probably going to get a lot more realistic, and at the same time, probably perform better.

Nice! So what's the solution for concave wings? Multiple colliders or custom definitions?

[Gfurst]

I'm sorry for ferram4, for he must suffer the most about complaints of his mods, things not working and people trying to learn plane design in this thread.

Maybe we should pull up another thread specifically for learning plane design and how to use FAR as whole and fly, leave the main thread for issues and fixes.

Well on another subject, I have a plane design question of my own.

I always find that my crafts glide for waaaaayyy to long without loosing speed. Ok its a good design, but seriously getting to kill speed with this a harder than anything.

Specially if I got into over supersonic speed, I just cut the throttle and sit there waiting for speed to drop while still not descending...

[Jovus]

Gfurst, that's not a design problem, that's the way air works.

If you want to slow down faster, you're going to need to do something to increase drag. Turning or pitching up are perennial favorites.

[ferram4]

@Van Disaster: I'd blame the biplane stuff as well. Even on traditional biplanes, you don't want the wings that close together.

@ThorBeon: Using parts with incorrect FARControlSurface modules. Nothing I can do, go talk to the mod author.

@legolegs: Noted, should be gone once the wing overhaul is finished. Until then, you'll have to deal with it.

@blowfish: The solution is custom definitions for concave wings. Simple solution for a simple problem.

@Gfurst: Yeah, surprisingly with proper aerodynamics slowing down is just as much of a pain as speeding up. Perhaps you're high enough up that the air density is way too low for drag to be that significant.

[Van Disaster]

@Gfurst: Altitude? cross section? I seem to remember those sort of complaints about a year ago. I've noticed it takes a long time to slow down re-entering, but the thicker part of the atmosphere is not somewhere you can glide forever for sure. Airliners ( which are resonably fat things ) have a glide ratio of about 18:1, what sort of glide ratio are you getting?

Ferram is the one who knows the mod details best, so if there's an odd quirk that's not obvious ( like my odd pitch-up thing which I've still no real idea about ) or the SPH looks like it's being odd then this really is the right place to go. A vehicle design thread is a nice idea, but that leaves the mod thread full of complaints

[Wanderfound]

I'm sorry for ferram4, for he must suffer the most about complaints of his mods, things not working and people trying to learn plane design in this thread.

Maybe we should pull up another thread specifically for learning plane design and how to use FAR as whole and fly, leave the main thread for issues and fixes.

Well on another subject, I have a plane design question of my own.

I always find that my crafts glide for waaaaayyy to long without loosing speed. Ok its a good design, but seriously getting to kill speed with this a harder than anything.

Specially if I got into over supersonic speed, I just cut the throttle and sit there waiting for speed to drop while still not descending...

A gliding plane is otherwise known as a falling plane. Slowing down rapidly is not something they do on their own.

You need to dissipate the energy of the aircraft somehow; S-turns, climb/dive cycles, low altitude drag, airbrakes (i.e. opposed spoilers and flaps). See the piloting guide in post #2 of the Kerbodyne thread linked below.

[Mekan1k]

So, just an update for people since I made a serious breakthrough on something I've wanted FAR to have for a long time. As you probably know, FAR requires wing geometry to be manually defined in a config file in order for it to handle everything properly. Further, it requires that every wing be able to be defined as a trapezoid, with the root and tip of the part being parallel. This, of course, is very limiting and frustrating, as well as a possible source of errors if someone implements it wrong.

Well, I just managed to implement an algorithm that allows FAR to determine the geometry of a wing from the mesh alone, assuming that it is completely convex in shape:

http://i.imgur.com/WZiVG6X.png

Now, this means that wings with concave portions (such as the stock old unmoving winglet and the gigantic B9 swept wings) won't be perfectly resolved with this. However, this system also allows manual definition of the shape for those wings, so stranger shapes can be accounted for.

Hopefully, more good news in the coming weeks on other stuff. FAR is probably going to get a lot more realistic, and at the same time, probably perform better.

ferram4, you are a modding God.

I didn't want to transfer to FAR because I didn't know how to apply the equations to mod parts that did not expressly have the values, but with this.... This I like.

Out of curiosity, if you have multiple shapes for a wing (such as that B9Expansion mod that existed a while ago), and switched between the shapes, would your new code be able to adapt to the new shape? Like for swing-wings and such?

[HoneyFox]

Hi ferram.

I've just sent a pull request to your git repo. See if you feel interested in those ideas.

Edited November 8, 2014 by HoneyFox

[Guest]

So, just an update for people since I made a serious breakthrough on something I've wanted FAR to have for a long time. As you probably know, FAR requires wing geometry to be manually defined in a config file in order for it to handle everything properly. Further, it requires that every wing be able to be defined as a trapezoid, with the root and tip of the part being parallel. This, of course, is very limiting and frustrating, as well as a possible source of errors if someone implements it wrong.

Well, I just managed to implement an algorithm that allows FAR to determine the geometry of a wing from the mesh alone, assuming that it is completely convex in shape:

http://i.imgur.com/WZiVG6X.png

Now, this means that wings with concave portions (such as the stock old unmoving winglet and the gigantic B9 swept wings) won't be perfectly resolved with this. However, this system also allows manual definition of the shape for those wings, so stranger shapes can be accounted for.

Hopefully, more good news in the coming weeks on other stuff. FAR is probably going to get a lot more realistic, and at the same time, probably perform better.

I want this. Really, really, really want this. That would solve a lot of problems, does that mean explicit FAR configuration would become obsolete on all but the weirdest shaped wings?